Configuring metrics on Server 2.16

⚠  CircleCI Server 2.x is no longer a supported release  ⚠
This article is legacy content

To configure AWS Cloudwatch, Datadog and Custom Metrics on the Server 2.16 installation, you will need to add config files for the metric provider.

You can create a file containing all configs for the output providers named *.conf and place it in /etc/circleconfig/telegraf/ like the following:

  - name: custom_telegraf_config_description
    type: label
    title: |
      Files matching `/etc/circleconfig/telegraf/*.conf` on this host will be included with the telegraf configuration.
      This allows you to specify custom output providers. For more information visit https://circleci.com/docs/2.0/monitoring/.
  - name: custom_telegraf_config_example
    title: Example
    type: textarea
    readonly: true
    value: |
      # SSH into the services host, then add the following to /etc/circleconfig/telegraf/kafka.conf
      # Afterwards restart the telegraf container with `sudo docker restart telegraf`
      [[outputs.kafka]]
        brokers = ["example.com:9092"]
        topic = "circleci"

Alternatively, you can create provider-specific configurations. In this example, we are collecting Datadog metrics and will name the file datadog.conf and place it in /etc/circleconfig/telegraf/. The file would look like the following:

[[outputs.datadog]]
  apikey = REPLACE_WITH_DATA_DOG_API_KEY

[[outputs.cloudwatch]]
  access_key = ""
  secret_key = ""
  region = ""
  namespace = "" 

Additional output providers can be found on the telegraph repo.